The buffers merge_arrow works in, kept by each thread from one join to the next (thread_arrow_workspace): mapping their pages anew took 11–16 % of an svd of order 64 to 512 on one x86-64 core, 13–20 % on six.
More...
#include <svd.hpp>
The buffers merge_arrow works in, kept by each thread from one join to the next (thread_arrow_workspace): mapping their pages anew took 11–16 % of an svd of order 64 to 512 on one x86-64 core, 13–20 % on six.
Definition at line 933 of file svd.hpp.
◆ above
poles[j] + roots[i]; then the right vectors.
Definition at line 936 of file svd.hpp.
◆ below
poles[j] − roots[i], row by root; then their product with above, then the left vectors.
Definition at line 935 of file svd.hpp.
◆ columns
One segment's rows of the kept columns that reach it.
Definition at line 939 of file svd.hpp.
◆ factor
The rows of the arrow's vectors for those columns.
Definition at line 940 of file svd.hpp.
◆ new_u
The kept columns of u times the left vectors.
Definition at line 937 of file svd.hpp.
◆ new_v
The kept columns of v times the right vectors.
Definition at line 938 of file svd.hpp.
The documentation for this struct was generated from the following file: